Вы находитесь на странице: 1из 10

2^3 Full Factorial Array (Similar to the Taguchi L8 Orthogonal Array)

By Kim Niles. (See accompanying article at http://www.isixsigma.com/library/bio/kniles.asp)


2 levels of 3 factors with up to 8 reps = (2^3) x 8 = 64 or less data points; Change only yellow boxes

Factor Assignment Response Table (Data) Quick Calculations


Main Effects Interactions (i.e. AxB). Replicates used to normalize the data
Run Order

G=ABC
D=AB
Run#

E=AC
F=BC
Signal-
to-Noise
A B C R1 R2 R3 R4 R5 R6 R7 R8 Ave. Std.Dev. (S/N)
1 - - - + + + - - - -
2 + - - - - + + - - -
3 - + - - + - + - - -
4 + + - + - - - - - -
5 - - + + - - + - - -
6 + - + - + - - - - -
7 - + + - - + - - - -
8 + + + + + + + - - -

Reference Key to Factor Selection


Factor A = Your factor: ;B= ;C=
Factor D = Your factor: AB Interaction? ; E = AC Interaction? ;F= BC Interaction? ; G = Error

Confounding Column Information (for alternative factor assignment considerations)


The Taguchi L8 array is very similar to this arrray but has slightly different columns that confound with each other. This
array does not have confounding columns. Using the L8 array requires an uderstanding of column confounding.

Calculating the Significance of Main Effects (Solving for Means)

Factor A Runs Where A is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where A is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor B Runs Where B is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where B is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor C Runs Where C is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where C is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Calculating the Significance of Main Effects (Solving for Standard Deviation)

Factor A Runs Where A is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where A is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor B Runs Where B is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where B is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor C Runs Where C is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where C is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Calculating the Significance of Main Effects (Solving for Signal-to-Noise)

Factor A Runs Where A is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where A is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

01/10/2011 23:54:58 49327496.xls 2^3 KN


Factor B Runs Where B is + : - + - + - + - = 0 / 4 = Ave= 0
Runs Where B is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor C Runs Where C is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where C is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Calc. the Sig. of 2nd Order Interaction Effects or Aliased Effects (Solving for Means)
Factor D (AB) Runs Where D is + : - + - + - + - = 0 / 4 = Ave= 0
Runs Where D is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor E (AC) Runs Where E is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where E is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor F (BC) Runs Where F is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where F is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Calc. the Sig. of 2nd Order Interaction Effects or Aliased Effects (Solving for Std Dev.)

Factor D (AB) Runs Where D is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where D is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor E (AC) Runs Where E is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where E is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor F (BC) Runs Where F is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where F is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Calc. the Sig. of 2nd Order Interaction Effects or Aliased Effects (Solving for Sig-to-Noise)

Factor D (AB) Runs Where D is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where D is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor E (AC) Runs Where E is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where E is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Factor F (BC) Runs Where F is + : - + - + - + - = 0 / 4 = Ave= 0


Runs Where F is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Calc. the Sig. of 3rd Order Interaction Effects, Error, or Aliased Effects (Solving for Means)
Factor G (ABC) Runs Where G is + : - + - + - + - = 0 / 4 = Ave= 0
Runs Where G is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Calc. the Sig. of 3rd Order Interaction Effects, Error, or Aliased Effects (Solving for Std.Dev.)

01/10/2011 23:55:00 49327496.xls 2^3 KN


Factor G (ABC) Runs Where G is + : - + - + - + - = 0 / 4 = Ave= 0
Runs Where G is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Calc. the Sig. of 3rd Order Interaction Effects, Error, or Aliased Effects (Solving for Sig.-to-Noise)
Factor G (ABC) Runs Where G is + : - + - + - + - = 0 / 4 = Ave= 0
Runs Where G is - : - + - + - + - = 0 / 4 = Ave= 0.00
Total Effect = Difference = 0.00

Calculating Confidence
ANOVA analysis is used to compare "within variation" with "between variation".
Hypotheses: Ho= All means are equal; HA= They are not equal
Linear contrasts is a method used to test any two levels or combinations of two levels against each other.
Degrees of freedom is always 1 for testing two items.

Calculating the Grand Average


Row Number: 1 2 3 4 5 6 7 8 Grand Average (GA)
Row Averages: - - - - - - - - => -

Factor Sum of Squares


(Linear Contrast Effect)^2/((1/n low level)+ (1/n high level)); n = number of data points
Factor A = 0.0 ^2 / (1/2) = 0.0
Factor B = 0.0 ^2 / (1/2) = 0.0
Factor C = 0.0 ^2 / (1/2) = 0.0
Factor D (AB) = 0.0 ^2 / (1/2) = 0.0
Factor E (AC) = 0.0 ^2 / (1/2) = 0
Factor F (BC) = 0.0 ^2 / (1/2) = 0.0
Factor G (ABC) = 0.0 ^2 / (1/2) = 0.0

Total Sum of Squares


Row Number: 1 2 3 4 5 6 7 8
Row Averages: - - - - - - - -
GA - Row Average: - - - - - - - -
(GA - Row Average)^2: - - - - - - - - Total SS Groups
Sum of (GA - Row Ave)^2 = 0.00

ANOVA Table
Is there a statistically significant difference between factors, rows, and error?
Df = Degrees of freedom MS = Mean Sum of Squares
SS = Sum of Squares F = F-test characteristic
Alpha = F-test probability

SS Df MS F Alpha Conf.
Factor A 0.0 1 0.00 - - - ≅ confidence that this is significant
Factor B 0.0 1 0.00 - - - ≅ confidence that this is significant
Factor C 0.0 1 0.00 - - - ≅ confidence that this is significant
Factor D (AB) 0.0 1 0.00 - - - ≅ confidence that this is significant
Factor E (AC) 0.0 1 0.00 - - - ≅ confidence that this is significant
Factor F (BC) 0.0 1 0.00 - - - ≅ confidence that this is significant
Factor G (ABC) 0.0 1 0.00
Total 0.0 7

Additional Comments
Note #1: If the data is known to be normally distributed, analyzing the within run data as design replicates yields
higher confidence in the results. Unfortunately this matrix is not set up to do that.
Note #2: Error in the results increases as columns contain greater amounts of partial data. For example, if two reps
are taken for each run, then columns R1 and R2 should be fully filled out.

01/10/2011 23:55:01 49327496.xls 2^3 KN


2^3 Full Fact. P
Factors (
Runs

C
A
B
1 + + + +
2 + + - +
3 + - + -
4 + - - -
http://www.isixsigma.com/index.php?option=com_k2&view=item&id=350:teaching-doe-using-formula-fuelers-toy-m

http://www.isixsigma.com/index.php?option=com_k2&view=itemlist&layout=category&task=category&id=51&Itemid

http://www.smartdraw.com/downloads/
sing-formula-fuelers-toy-motorcycles&Itemid=156

k=category&id=51&Itemid=184
Generic Steps For Using The Attached Spreadsheet
There are many different articles in the literature that outline steps that should be taken to complete a DOE. The following steps

1. Determine the acceptance criteria you need (i.e. acceptable alpha error or confidence level for determining what you will acc
solid decisions; see additional advice below.

2. Pick 2-3 factors to be tested and assign them to columns A, B, and C as applicable (advise using the key provided).

3. Pick 2 different test levels for each of the factors you picked (i.e. low/high, on/off, etc.).

4. Determine the number of samples per run (room for 1-8 only; affects normality and effect accuracy, not confidence).

5. Randomize the order to the extent possible.

6. Run the experiment and collect data. Keep track of everything you think could be important (i.e. people, material lot numbers
these may affect the validity of the conclusions.

7. Analyze the data by entering the data into the yellow boxes of the spreadsheet and reading the results. A review of the ANO
established in step number one. If the alpha value in the table is greater than the acceptance criteria, accept the result; if it is le
probability that that factor is statistically different from the others. Signal to noise measurements are helpful to use when selecti

8. Confirm your results by performing a separate test, another DOE, or in some other way before fully accepting any results. Yo
criteria by retesting the factor using larger differences between the levels.
Additional Advice
How DOEs Work
Note that using the eight run array, we have four runs being tested with each factor at high levels and four without being at a hi
of each high level (4 high + 4 not high = 8 relative to high) and vice versa for each factor and the interactions between the three
test becomes the statistical equivalent of a 96 run, one-factor-at-a-time (OFAT) test [(8 Ahigh)+(8 Alow)+(8 Bhigh)+(8 Blow)+(8
+(8 BClow)]. Other advantages to using DOE include the ability to use statistical software to make predictions about any combi
generating various types of informative two- and three-dimensional plots. Therefore, DOEs produce orders of magnitude more

DOEs don't directly compare results against a control or standard. They evaluate all effects and interactions and determine if th
statistical confidence levels for each measurement. A large effect might result but if the statistical confidence for that measurem
measured effect with high confidence tells us that the effect really isn't important.

Why This Array?


Selecting a test array requires balancing your test objectives, test conditions, test strategy, and resources available. Therefore,
factors at once than one large DOE. Comparing the statistical power of this array (inherent ability to resolve differences betwee
runs needed) also shows how this array is advantageous since it requires only eight runs and yields successful results in most

Picking Acceptance Criteria


Acceptable confidence depends upon your needs. If health could be affected, then you may want more than 99% confidence b
considered low but in some cases is worth considering (see precautions below). A confidence level of 80-90% is considered mo
level greater than 90% is considered high with the results usually considered very likely to be correct.

Calculating Samples Per Run


Since statistical confidence doesn't increase with additional samples per run (replications are needed to have that effect), it is im
concerns of non-normal data exist (in accordance with the central limit theorem) and or to improve measured effect accuracy (s
the statistical confidence in our effects is usually more important than statistical significance and normality effects, running two

Precautions
Since this array has less power than others available, we need to remember that when optimizing a process that isn't critical to
much better than not knowing which way to go with machine settings, etc. Assuming all the error in one's experiment is evenly
(measured to be true via DOE), for example, might seem horrible but really means the equivalent of 80%, since the 40% that w

The accompanying spreadsheet cannot easily be changed. It should be used while training others (shows the math), or when y
software. It can't be replicated or you can't add center points in its current form (center points increase statistical confidence by

About the Author: Kim Niles has more than 17 years process control and improvement experience working with San Diego ma
Currently an officer in three professional societies, Mr. Niles has a master's degree in quality science with an emphasis in Six S
degree from San Diego State University through the industrial technology department. He can be reached at kim.niles@cox.ne
JavaScript enabled to view it .

Вам также может понравиться